Abstract
We investigate the use of particle filter (PF) estimation techniques on a hovercraft vehicle in an office environment. Monte Carlo Localization (MCL) with particle filtering is a popular method for localizing robots with laser range finders. In maps featuring long, uniform corridors though, a PF can produce low confidence estimates. When used as feedback to control an unstable vehicle this can prove fatal. This is because, unlike grounded wheeled vehicles, an airborne hovercraft requires accurate localization
not only for path planning, but for stabilization as well. We solve the low confidence problem using a secondary networked robot as a mobile map feature.
not only for path planning, but for stabilization as well. We solve the low confidence problem using a secondary networked robot as a mobile map feature.
| Original language | English |
|---|---|
| Publication status | Published - 2009 |
| Event | Fourth Swedish Workshop on Autonomous Robotics SWAR'09 - Västerås, Sweden Duration: 2009 Sept 1 → … |
Conference
| Conference | Fourth Swedish Workshop on Autonomous Robotics SWAR'09 |
|---|---|
| Country/Territory | Sweden |
| City | Västerås |
| Period | 2009/09/01 → … |
Subject classification (UKÄ)
- Control Engineering
Free keywords
- localization
- particle filter
- hovercraft
- networked autonomous vehicles
- laser range finder